Can we believe the hype? Let’s look at generative AI in gaming

The gaming industry has long been renowned for ongoing and rapid transformation. Yet, this is also an industry that has not been exempt from recent headlines focused on artificial intelligence. Like we've seen across many other sectors, generative AI is now coming for gaming. Or is it? 

Applications such as Lensa, Mid-Journey and, of course, ChatGPT have dominated the news cycle recently, permeating our collective consciousness and sending ripples across the internet. Each app has driven users into a frenzy, each having been purported to have the very real potential to create visual and written art better than humans – definitively replacing them in the workplace. 

But where did this sudden global interest originate? The word on the street is that, in line with the cryptocurrency bubble bursting, those involved in its initial ascension have now shifted focus to generative AI; pitching it as the next big thing. These timelines certainly align with the rise of popular applications. In November, Lensa AI (among others) gathered a lot of attention, as users began uploading images awash with errors and accusations of plagiarism to social media platforms. While the flurry of interest was certainly immense, it was ultimately met with negative backlash and a class action lawsuit. 

Yet, it’s a phenomenon transforming conversations across the content creation landscape – and the frenzy to automate artist and creator work doesn’t look ready to slow down (we’re looking at you, ChatGPT). In fact, the hype would lead audiences to believe that progress really does lie in automation. But as far as the role in generative AI in gaming goes – can the hype live up to expectations? 

How is generative AI set to impact gaming? 

In short, generative AI needs to play the long-game. Artificial intelligence is already ubiquitous in gaming, and the current collective misunderstanding around the role it plays needs to be assessed before we see an impactful change. 

Take the proliferation of open-world gaming as an example. Games such as The Elder Scrolls V: Skyrim and Grand Theft Auto have long-relied on the pre-programmed behaviour of non-playable characters (NPCs) to bring diversity to in-game interactions. It seems only natural that gamers would anticipate a ChatGPT type of integration to avoid specific interactions becoming too repetitive when roaming the plains of Tamriel. 

While text-based generative AI will have a role in gaming eventually, right now we can’t assume that this technology is a silver-bullet to alleviate gamer woes. The work required to deepen NPC conversations and create organic in-game dynamics between players and NPCs is extensive. Right now, generative AI would need to be custom built to live up to the hype, and even then, this would require human supervision and oversight. 

There will come a day where developers can give characters personality traits that allow generative AI to take the wheel and create behaviour indistinguishable from the real-life equivalent, but today, it has its limitations. NPCs based on current models have the potential to cause trouble. Not to linger too long on Bing’s recent mishap, but current iterations of text-based generative AI could go off on tangents and ultimately, discuss topics outside of the game. In a game like Grand Theft Auto, likely not a huge issue, but for a game like Skyrim, conversations beyond the lore could prove problematic. 

What’s more, games that haven’t been created with AI in mind can’t simply have generative AI dropped in. This is more likely to cause a wealth of issues that require human input to resolve. Certain games would require a total redesign to accommodate this type of tech, and in an industry with huge budgets and tight margins, that simply isn’t realistic right now. 

Generative AI has already found a place in the world of gaming 

Despite its infancy, there is no denying that ChatGPT is a disruptive piece of technology. Its impact has triggered a monumental paradigm shift for the whole world. Its potential is massive – far greater than current scenes being played out between Microsoft, Google and their respective search engines – that’s for sure. While generative AI does require a lot of work, there are some solid examples of how it is being utilised currently, ensuring that the future does, indeed, look bright for generative AI. 

OpenAI – the company behind ChatGPT – have already demonstrated the potential of artificial intelligence and deep reinforcement learning algorithms. Building on the success achieved by Google's Deepmind, with complex board games like Chess and Go, OpenAI pioneered reinforcement learning agents - a type of AI that learns to make decisions based on environmental feedback. This demonstrated the true potential of AI and deep reinforcement learning algorithms in complex real-time strategy games. In the complex, multiplayer online battle arena (MOBA) real-time strategy game Dota 2, this technology was pitted against the world’s top professional players in both 1v1 and team matches. Crushingly, it won 99.4% of the time. 

Microsoft’s Flight Simulator has utilised Blackshark.ai to generate a photorealistic 3D world from 2D satellite images to make the in-game flying environment more realistic. Using real depth of field across an entire world is a huge feat to scale – this would have been impossible to accomplish without AI. 

The future of generative AI in gaming looks bright, though!  

While generative AI won’t have a near-term disruptive impact on the gaming industry, the progression and adoption of this type of technology will certainly have an impact beyond ideation in design and programming, for example. 

Indeed, Roblox is bringing generative AI to its gaming universe in the form of a tool that will help produce Roblox computer code for an in-game interaction from a short text prompt. This tool is particularly relevant – and vital – for the Roblox player demographic, which is predominantly children. Lowering the barrier to entry for coding, this educational tool and AI integration will enable children to pour their creativity into the gaming world, while teaching them important developmental skills. 

What’s to come? 

As we’ve learned from the gaming industry in the last few decades, it’s one almost perpetually on the cusp of disruption. With the introduction of ChatGPT and other generative AI tools aiming to fundamentally shift how we live, work and play, there is no doubt that generative AI will have a crucial role in transforming the gaming industry. 

Harking back to NPCs, it doesn’t feel out of the realms of possibility for these non-playable characters to change their personality based on player actions, or develop new goals infinitely and independently and even build rapport with a player over time. This would conceive a genuine attachment to NPCs as well as a player’s ability to interact with various good/evil characters, enriching the in-game experience. 

Of course, the audio/speech work for this would be the most challenging step, but there are already a plethora of tools that exist to simulate a real person’s speech. In game, this would manifest itself as an entire world existing outside of the player; progressive side quests that don’t appear scripted, quests that are determined by a player’s actions in-game, with NPC characteristics defining and influencing their judgement of what a player says/does in-game. When combined with virtual reality, this is a recipe for an ultra-realistic environment, where in-game interactions mimic those of real-life – perhaps a terrifying notion for some! 

This could also lend itself to competitive gaming – a new era of sophisticated matchmaking systems that can track a wide range of factors, with a much deeper granularity that goes far beyond general ranking. This could look at skill level and play-style, tracked via in-game movements and preferences for certain item purchases defining a match. As players choose to challenge themselves more with such matchmaking, even casual gamers will be able to, inevitably, evolve their gaming skills through experiential learning. 

Of course, as Microsoft continues its mission to integrate ChatGPT tech across its fleet of services, one can’t help but wonder what GamePass will look like with such an upgrade. With improved user experience, games matching can become more personalised, games built in-house are likely to come with an in-game virtual assistant feature – think Jarvis in Iron Man or Glados in Portal – that can be leveraged to perform tasks in-game without having to pause or go into menus. 

Indeed, a tool like this could be a powerful bargaining chip when signing a new studio. As Microsoft’s game matching becomes more sophisticated, it makes commercial sense for Microsoft to load up on independent titles that suit every gamer palette and elevating player experience in games with more immersive storytelling and characters. A feature many independent studios would benefit from is scripted writing and design prototyping. Taking the weight of development from smaller teams will allow for more creativity and new game concepts to emerge. 

While it may be a little too early to assume generative AI is an omnipotent technology that will turn the industry on its head, the potential of these tools currently encouraging us all to hop aboard the hype train is exciting. In an industry that is bound by absolutely nothing when it comes to creativity, innovation and potential, the future of generative AI burns bright. 

For enquiries about our gaming content, please reach out to our press liaison at: vicki.hayward@futuresource-hq.com

About the author

Arian Bassari

Arian joined the Gaming team at Futuresource in 2021 as an analyst, where he conducts research to understand market dynamics and provide data-driven insights.